· · · Repository hosted by TU Delft Library

Skin detection is one of the building blocks of many image processing and computer vision algorithms. Within the Video Processing group at Philips Research we are investigating the use of skin detection algorithms for a number of core applications. In this report we focus on skin color based methods. These methods are extremely popular in the computer vision and image processing fields because they are simple and fast, they are robust to geometric variations of the skinpatterns, to partial occlusions, scale or rotation changes. The aim of this report is to discuss the different aspects involved in the implementation of a pixel-based skin detection algorithm. These questions will drive the practical realization of the system, which will be discussed and evaluated at the end of the manuscript.

This report describes the activities carried out in the first part of the Experiencing Control project (2008-324). The guiding idea of the project is to make control part of the experience, exploring new interaction solutions for complex, engaging interactions with Philips devices in the living room. In particular we focus on a collaborative content creation scenario, for which a concept prototype, the Q-ball, has been build and demonstrated.

We consider the problem of indoor user zoning - identifying occupiedspatial regions by users in a room, and tracking using a grid of passive infrared (PIR) sensors. Under the proposed scheme, quantized levels of analog signal output from pairs of PIR sensors are used todetermine an occupied zone. Viterbi-based tracking is employed to further improve the robustness of zoning and track the zones traversedby a user. Experiment results are presented to validate the efficacy of the proposed method for one-dimensional user zoning and tracking.

This document describes the results of a technology survey for device-less interaction. The Device-less Interaction project (2007-307) aims at providing interaction options for future home appliances without resorting to a remote control or any other dedicated control device. The target home appliances that we have in mind are audio and audio-visual devices, with an emphasis on audio use cases. Indeed, device-less interaction can create a strong differentiator for a series of AVM products in key application scenarios. This survey is essentially focused on the technologies and not on the devices that should embed such technologies. In this sense thus the survey is general and not restricted to these use cases. Still, we will also discussin this document relevant device-less interaction use cases for BU AVM and use these use cases as examples to validate the technology landscape built here.

There is a growing interest in multi-modal signal processing: sets of related signals are jointly processed to extract information that is otherwise hidden when considering the different modalities independently. One popular problem in crossmodal processing is the localization of visual sources synchronous with audio stimuli. Audiovisual source localization allows to pinpoint and extract salient audio-video information from a scene, enabling innovative applications in communication, interaction and gaming. In this paper we aim to achieve cross-modal localization in real-time using single camera, single microphone data. Existing works use complex statistical data models or complex representations of audio and video features, limiting their applicability in realtime systems. In this paper we propose a simple yet effective algorithm that allows to detect and localize in real-time synchronous audio-video sources. The proposed approach obtains the best speaker localization performances reported to date on the popular CUAVE database, while running in real-time and without requiring any training.

While it has been widely proved that breastfeeding is the healthiestfeeding option for a baby and its mother, the mechanisms by which ababy removes milk from the breast are still not completely known. Partly this is due to the lack of tools to analyze images of the infant oral cavity during feeding automatically and quantitatively. In this paper we propose two methods for analyzing ultrasound videos toautomatically detect relevant events such sucking and swallowing ofmilk and to discriminate different types of tongue action during milk removal. The proposed algorithms provide, for the first time, quantitative indications of the type of activities carried out during breastfeeding by the baby, promising unprecedented advancements in thefield.

LumaFluid is an interactive environment that explores new ways to stimulate emotional and social engagement through immersive light effects. A computer vision system detects and tracks persons present inthe LumaFluid square. Using this location information, colored spotlights highlight each person on the installation floor. Connections are also drawn between neighboring visitors, so that dynamic, colorful effects underline and stimulate interpersonal communication. Two versions of the concept where deployed during ten days at the 2011 STRP Art and Technology Festival. In this paper we describe the motivation and the realization of the installation, and we discuss insights and reactions collected during the event.